Matches in SemOpenAlex for { <https://semopenalex.org/work/W3162192368> ?p ?o ?g. }
- W3162192368 endingPage "107129" @default.
- W3162192368 startingPage "107129" @default.
- W3162192368 abstract "Few-shot learning aims to train an effective classifier in a small data regime. Due to the scarcity of training samples (usually as small as 1 or 5), traditional deep learning solutions often suffer from overfitting. To address this issue, an intuitive idea is to augment or hallucinate sufficient training data. For this purpose, in this paper, we propose a simple yet effective method to build a model for novel categories with few samples. Specifically, we assume that each category in the base set follows a Gaussian distribution, so that we can employ Maximum A Posteriori (MAP) to estimate the distribution of a novel category with even one example. To achieve this goal, we first transform each base category into Gaussian form with power transformation for MAP estimation. Then, we estimate the Gaussian mean of the novel category under the Gaussian prior given few samples from it. Finally, each novel category is represented by a unique Gaussian distribution, where sufficient trainable features can be sampled to obtain a highly accurate classifier for final predictions. Experimental results on four few-shot benchmarks show that it significantly outperforms the baseline methods on both 1- and 5-shot tasks. Extensive results on cross-domain tasks and visualization of estimated feature distribution also demonstrate its effectiveness." @default.
- W3162192368 created "2021-05-24" @default.
- W3162192368 creator A5006191181 @default.
- W3162192368 creator A5008056956 @default.
- W3162192368 creator A5018598136 @default.
- W3162192368 creator A5028369180 @default.
- W3162192368 creator A5044961935 @default.
- W3162192368 date "2021-08-01" @default.
- W3162192368 modified "2023-10-05" @default.
- W3162192368 title "Feature hallucination via Maximum A Posteriori for few-shot learning" @default.
- W3162192368 cites W1498436455 @default.
- W3162192368 cites W1677182931 @default.
- W3162192368 cites W2194321275 @default.
- W3162192368 cites W2832876791 @default.
- W3162192368 cites W2904218366 @default.
- W3162192368 cites W2909383239 @default.
- W3162192368 cites W2949879676 @default.
- W3162192368 cites W2962799101 @default.
- W3162192368 cites W2963070905 @default.
- W3162192368 cites W2963078860 @default.
- W3162192368 cites W2963350370 @default.
- W3162192368 cites W2963845150 @default.
- W3162192368 cites W2963960318 @default.
- W3162192368 cites W2964105864 @default.
- W3162192368 cites W2967333288 @default.
- W3162192368 cites W2988205463 @default.
- W3162192368 cites W2988501586 @default.
- W3162192368 cites W2990216037 @default.
- W3162192368 cites W3001807478 @default.
- W3162192368 cites W3009081299 @default.
- W3162192368 cites W3012255272 @default.
- W3162192368 cites W3034187513 @default.
- W3162192368 cites W3034978645 @default.
- W3162192368 cites W3035143213 @default.
- W3162192368 cites W3035163205 @default.
- W3162192368 cites W3035370595 @default.
- W3162192368 cites W3088750566 @default.
- W3162192368 cites W3095388829 @default.
- W3162192368 cites W3099486271 @default.
- W3162192368 cites W3103819522 @default.
- W3162192368 cites W3126133447 @default.
- W3162192368 doi "https://doi.org/10.1016/j.knosys.2021.107129" @default.
- W3162192368 hasPublicationYear "2021" @default.
- W3162192368 type Work @default.
- W3162192368 sameAs 3162192368 @default.
- W3162192368 citedByCount "11" @default.
- W3162192368 countsByYear W31621923682022 @default.
- W3162192368 countsByYear W31621923682023 @default.
- W3162192368 crossrefType "journal-article" @default.
- W3162192368 hasAuthorship W3162192368A5006191181 @default.
- W3162192368 hasAuthorship W3162192368A5008056956 @default.
- W3162192368 hasAuthorship W3162192368A5018598136 @default.
- W3162192368 hasAuthorship W3162192368A5028369180 @default.
- W3162192368 hasAuthorship W3162192368A5044961935 @default.
- W3162192368 hasBestOaLocation W31621923681 @default.
- W3162192368 hasConcept C105795698 @default.
- W3162192368 hasConcept C111472728 @default.
- W3162192368 hasConcept C119857082 @default.
- W3162192368 hasConcept C121332964 @default.
- W3162192368 hasConcept C138885662 @default.
- W3162192368 hasConcept C153180895 @default.
- W3162192368 hasConcept C154945302 @default.
- W3162192368 hasConcept C163716315 @default.
- W3162192368 hasConcept C22019652 @default.
- W3162192368 hasConcept C2911011789 @default.
- W3162192368 hasConcept C33923547 @default.
- W3162192368 hasConcept C41008148 @default.
- W3162192368 hasConcept C49781872 @default.
- W3162192368 hasConcept C50644808 @default.
- W3162192368 hasConcept C62520636 @default.
- W3162192368 hasConcept C75553542 @default.
- W3162192368 hasConcept C95623464 @default.
- W3162192368 hasConcept C9810830 @default.
- W3162192368 hasConceptScore W3162192368C105795698 @default.
- W3162192368 hasConceptScore W3162192368C111472728 @default.
- W3162192368 hasConceptScore W3162192368C119857082 @default.
- W3162192368 hasConceptScore W3162192368C121332964 @default.
- W3162192368 hasConceptScore W3162192368C138885662 @default.
- W3162192368 hasConceptScore W3162192368C153180895 @default.
- W3162192368 hasConceptScore W3162192368C154945302 @default.
- W3162192368 hasConceptScore W3162192368C163716315 @default.
- W3162192368 hasConceptScore W3162192368C22019652 @default.
- W3162192368 hasConceptScore W3162192368C2911011789 @default.
- W3162192368 hasConceptScore W3162192368C33923547 @default.
- W3162192368 hasConceptScore W3162192368C41008148 @default.
- W3162192368 hasConceptScore W3162192368C49781872 @default.
- W3162192368 hasConceptScore W3162192368C50644808 @default.
- W3162192368 hasConceptScore W3162192368C62520636 @default.
- W3162192368 hasConceptScore W3162192368C75553542 @default.
- W3162192368 hasConceptScore W3162192368C95623464 @default.
- W3162192368 hasConceptScore W3162192368C9810830 @default.
- W3162192368 hasFunder F4320322725 @default.
- W3162192368 hasFunder F4320328119 @default.
- W3162192368 hasLocation W31621923681 @default.
- W3162192368 hasOpenAccess W3162192368 @default.
- W3162192368 hasPrimaryLocation W31621923681 @default.
- W3162192368 hasRelatedWork W1996541855 @default.
- W3162192368 hasRelatedWork W2563096758 @default.